Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Large Parameter Estimates for Price with a Small Sample


I have fielded an ACBC study and I'm getting quite large utility estimates for price. I've used summed price in the survey and this particular survey has 66 respondents which is relatively small - but should still be estimable. If I estimate the model between the highest and lowest prices seen by respondents (after doing a counts run and looking at the distribution of prices seen by respondents) I get values of between approximately +/-8  using either monotone regression of HB. Intuitively 8 seems to high to me, but I could be wrong. When I use pricewise estimation for price in HB and use a few cut points the difference between the estimates becomes much larger and if i reduce the range of prices the number of respondents drops (as the estimation only takes into account those respondents that saw prices within the range). However if I use monotone and set a few post hoc  price points then the parameter estimates become much more reasonable i.e. +/-3.

I guess my questions are, is a utility range between +/-8 too large? should I be using HB or Monotone in this instance? and is applying post hoc price points for a linear or loglinear price estimation with a more restrictive price range an appropriate thing to do given the distribution of prices and the size of the sample?
asked Jun 20, 2017 by Jasha Bowe Bronze (1,745 points)

Your answer

Please only use this to answer the original question. Otherwise please use comments.
Your name to display (optional):
Privacy: Your email address will only be used for sending these notifications.
Anti-spam verification:

To avoid this verification in future, please log in or register.